Work

RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations

Year: 2024
Type: conference-paper
Abstract: Individual neurons participate in the representation of multiple high-level concepts.To what extent can different interpretability methods successfully disentangle these roles?To help address this question, we introduce RAVEL (Resolving Attribute-Value Entanglements in Language Models), a dataset that enables tightly controlled, quantitative comparisons between a variety of existing interpretability methods.We use the resulting conceptual framework to define the new method of Multi-task Distributed Alignment Search (MDAS), which allows us to find distributed representations satisfying multiple causal criteria.With Llama2-7B as the target language model, MDAS achieves state-of-the-art results on RAVEL, demonstrating the importance of going beyond neuron-level analyses to identify features distributed across activations.We release our benchmark at https://github.com/ explanare/ravel.
Language: English
Open Access status: hybrid
OpenAlex ID: W4402671619
FWCI: 9.314
Cites: 13
Cited by: 61
Related to: 10